Phase Change Material Product Design. Market and Business Development Assessment in the Food Industry Davide Lora

Home , Vacuum flask

LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 1 Phase change material product design. Market and business development assessment in the food industry Davide Lora

Abstract poured in, it is then possible to identify three phases. The first Tea and coffee are very popular beverages but are often one is identified by the fact that since the water was boiling, served too hot and do not stay warm for long periods the liquid is too hot to drink. As the drink slowly cools down, of time. A phase change material (PCM) can be used it is possible to define the second phase as the time where the effectively as a passive heat storage method since it can drink keeps the optimum drinking temperature. The last phase absorb large quantities of latent heat. Preheating is not required since it absorbs heat at temperatures at which starts at the moment when the drink is regarded lukewarm the liquid cannot be drunk. The use of such materials until it progressively cools towards ambient temperature. The in combination with liquid containers (a thermos and result is a very limited time frame in which it is possible to a cup) is investigated. Two PCM materials, a paraffin enjoy any beverage at optimum temperature. The final user ◦ wax (Tfus=55,84 C) and sodium acetate trihydrate ◦ experience is not as enjoyable as it should be: the user initially (Tfus=58 C) are experimented. Results show that using these materials in a thermos, a liquid can be cooled to can get burnt, and after that, he or she has to rush in order to Optimum Temperature (topt=60 ◦C) up to 95 % faster and finish the drink before reaching lukewarm temperature. the optimum drinking temperature range, defined between 60 and 50 ◦C, can be kept around 50 % longer. For a cup, results are dependent on the intended use, but are A. Objectives generally discordant since the absence of proper thermal As concern arises around global warming and the impact isolation hinders the heat absorption process of the PCM. that each energy source has in creating electrical and thermal A business model and a macro and micro environment analysis are also performed and show a market with an energy, it is important to tackle this issues by increasing energy high interest towards these kind of products. Finally, a efficiency without providing or using additional energy. The possible design for the thermos and PCM containers is purpose of this work is to propose a solution through means of also proposed and is determined financially viable. passive energy storage methods by using phase change materials (PCMs). If this concept is applied to the earlier example of Index Terms—Phase change material, PCM, coffee, tea, thermos, vacuum flask, cup, mug, food industry. the thermos filled with coffee, it translates in an attempt of increasing the second phase time frame – the one associated to the optimal drinking temperature – while decreasing the time I.INTRODUCTION the liquid stays in the other two phases. S restaurants all over the world strive to offer customers A their best tasting experience, the temperature at which II.STATE OF THE ART food is served is often overlooked. As outlined by several To tackle these objectives, a PCM was employed. They studies [1], [2] temperature is a key factor in determining are materials that are able to store large quantities of heat in how humans taste food. As with food, also the temperature a relatively small mass and with small temperature changes. of the drinks affects our tasting experiences. Depending on In this chapter an overview of their working principles and the culture, a meal is usually accompanied by cold water, characteristics is given. ambient temperature water or tea in North America, Europe and Asia respectively [3]. A correlation between drinking water temperature and the perceived intensity and creaminess of A. Latent heat storage materials sweets was proven, resulting in a preference for North American Thermal energy can be stored using different physical people to highly sweetened foods [3]. In addition, temperature or chemical processes, and each method has quite different is often a discriminant for determining how good a drink is. characteristics. The heat stored to change the temperature Many beverages widely enjoyed across many different cultures, of any substance is called sensible heat [4]. A change in such as coffee and tea, are best savored when very hot. Others, stored energy corresponds to a direct temperature change. especially during summer, are required to be always drank In addition, heat can also be stored as latent heat, i.e. the cold. amount of heat absorbed or released during a phase change. In A conventional thermos uses a vacuum gap between two solid-liquid transformation, the phase change is an isothermal metal walls to accomplish this task. If an optimum drinking process. High amounts of heat can be stored during the phase temperature is defined and, for example, fresh brewed coffee is change while the temperature remains mostly constant. This LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 2 is a very important characteristic since each PCM presents a makes the liquids available at an optimum drinking temperature different melting temperature. Once the melting is completed, for longer periods of time, in confront of a standard thermos further supply of heat creates only an increase in sensible heat. resulting in energy savings. This particular behavior allows to choose a PCM depending on the application, by simply choosing the desired melting A. PCMs and optimal drinking temperature temperature. Additionally, it allows to have an accurate control of the temperature. The optimal drinking temperature was selected by research- PCMs present very different thermal and chemical properties. ing and analyzing other works. A company, already selling ◦ It is thus necessary to understand their main differences and a similar system, uses a PCM temperature of 60 C [7]. to categorize them in order to select the most performing Additional studies identified the optimal drinking temperature ◦ and proper material for any given application. In Figure 1 to around 57,8 C [8]. a classification method is proposed and the main properties, 1) Paraffin Wax: A commercial paraffin wax was chosen as advantages, disadvantages and examples, are reported for each the main PCM, since, as highlighted in the previous chapter, category of materials. Important properties that define the they provide a wide range of melting temperatures, relatively high latent heat of fusions, constant performance through characteristic of any PCM are the latent heat of fusion (∆Hfus) and the melting temperature (T ). Other properties such as cycling and are relatively inexpensive, as well as non toxic [6]. fus ◦ density (ρ), melting behavior and supercooling also need to be A paraffin wax with a nominal melting point between 53 C ◦ taken into consideration. to 57 C and produced by Sigma-Aldrich was chosen. Three different melting behaviors exist: congruent melting 2) Sodium Acetate Trihydrate: Most of the experiments where, given a specific melting temperature, the salt in its conducted during this work were developed using the paraffin anhydrous form is completely soluble in water and there is no wax. To test possible performance improvements given by a phase separation. Incongruent melting: is given when the salt in more dense and higher melting point PCM, a test with sodium its anhydrous form is not completely soluble in water resulting acetate trihydrate Na(CH3COO) · 3 H2O was performed. Table I in extensive phase separation. Semi-congruent melting presents reports the main properties of this material. an intermediate behavior between congruent and incongruent melting [5]. Table I PHYSICAL AND THERMAL PROPERTIES OF SODIUM ACETATE TRIHYDRATE Supercooling is a particular state of a material that is found [9], [10]. in its liquid form at a temperature below its freezing point [6]. ∗ ∗ ∗ Tfus ∆Hfus ρ cp k [kg m−3] [kJ kg−1 K−1] [W m−1 K−1] [◦C] [kJ kg−1] Phase Change Materials Solid Liquid Solid Liquid Solid Liquid 57−58,5 260 ±11 1450 1280 2,79 3,0 0,7 0,4 ∗: at 58 ◦C Paraffin Compounds Organic To suppress the supercooling behavior a nucleating agent Non-paraffin Compounds was used (disodium hydrogen phosphate dodecahydrate (DSP) Na2HPO4 · 12 H2O) [11]. A solution packed in a single double wall plastic bag was experimented using 124,1 g of sodium Salt Hydrate acetate trihydrate and 7,6 g of DSP. Inorganic Metallic B. Analysis and characterization 1) Differential scanning calorimetry: The machine used Organic - Organic is a TA DSC 2920 produced by TA Instruments. The mass of the sample used in the text was measured in a controlled Eutectic Inorganic - Inorganic environment with a Mettler Toledo UMT2 ultra-micro balance. A paraffin wax sample mass was measured and sealed inside an Inorganic - Organic aluminum pan and placed inside the furnace with a reference empty one. The sample first brought to a temperature of −10 ◦C, ◦ Figure 1. Classification of latent heat storage materials. then successively a heating cycle up to 80 C was performed with a rate of 10 ◦C min−1. After the first heating cycle, a cooling cycle brought the samples back to −10 ◦C and a second III.METHODOLOGY heating cycle was repeated. This last cooling and heating cycles where used to determine the properties of the paraffin wax. The use of a PCM material is investigated with the purpose of creating a system that is able to retain the thermal energy of water for a long period of time. The goal is to create a system C. Density that improves the current thermos technology and provides Density measurements were done using the following hot liquids (or cold beverages) in a faster way. In addition, it methodology. The paraffin wax was liquefied using a thermal LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 3 bath filled with a low viscosity silicon oil. A predetermined the water to PCM ratio can be defined considering that all the volume (500 µL) of paraffin wax was taken using a Rainin energy contained by the water (inserted at 90 ◦C) is absorbed by Pipet-lite LTS 100-1000uL (error: ±4µL). Once solidified, the the paraffin wax. The water specific heat capacity is considered −1 −1 mass was then measured using a Mettler Toledo Newclassic constant and equal to cp-w=4184 J kg K [4]. All the heat MF - MS205DU balance (error of ±10µg). The error on the contained by the water is absorbed by the paraffin wax (that temperature measurement of the thermal bath is of ±1 ◦C. is found at ambient temperature) as specific and latent heat. 1) Viscosity: To measure the viscosity a DV-II+ Pro viscosi- After a certain time, the water and the paraffin will be both at meter, produced by Brookfield Engineering Laboratories, was the PCM melting point temperature and all the paraffin wax used. The accuracy associated to the viscosimeter is of the will be completely melted. 1 %. Using a thermal bath, a predetermined mass of paraffin was inserted in the apparatus at 80 ◦C. The viscosimeter, was m ·c ·(90−T )+c ·(T −20) = m ·∆H able to regulate the temperature of the sample with a precision w p-w fus p-PCM fus PCM fus (1) ◦ of ±1 C. The kinetic viscosity was then calculated. Applying equation 1 to the paraffin wax and to the sodium 2) Laser flash analysis: Through the laser flash analysis acetate it was possible to obtain equation 2 and 3 respectively. (LFA) method it was possible to measure the thermal diffusivity For the paraffin wax the results obtained from the differential (α). The apparatus used is a Flashline 5000 from Anter. The scanning calorimeter (DSC) and reported in Figure 4 were used, accuracy of the machine is of ±3%. Several samples of different whereas for the sodium acetate trihydrate the values reported thicknesses and 0,5 inch (12,7 mm) diameter were prepared in Table I were employed. A general value for the specific beforehand since depending on the expected diffusivity a certain heat capacity was used [4]. thickness is required. The final thickness of the sample was of m 182, 4 + 2, 89 · (55, 84 − 20) 2 mm. Five shots with the laser were used for each temperature. w = = 2, 0 (2) The average value between these shots is the value considered mPCM 4, 184 · (90 − 55, 84) valid. m 260 + 2, 79 · (58 − 20) w = = 2, 73 (3) mPCM 4, 184 · (90 − 58) D. Datalogger and thermocouples With this data it is possible to know the minimum water An Omega OM-DAQPRO-5300 datalogger was used to to PCM mass ratio that is useful. Mass ratios below these record the temperatures. The accuracy provided is of ±0,5% values, and keeping all the other parameters constant, means ◦ ◦ for temperatures above 50 C and of ±0,5 C for temperatures inserting more PCM which results in a decrease in the amount ◦ ◦ between −50 C to 50 C. In addition, there is an error of of available water without fully exploiting the additional PCM ±0,3% on all measurements due to the thermocouples cold inserted. junction. 1) The reference experiment: The thermocouple position for Two different types of K type thermocouples produced a typical run of the reference experiment is shown in Figure 2 by Omega were used for all the experiments. A thin type (only T1,T2,T3,T4,T5 and T6 were used). ("KMQXL-IM050U-450") and a thicker type ("KMQXL- The procedures for a typical run would be the following: IM150U-450"). The first type was used only for T5 with first distilled water was heated at 90 ◦C using a thermal bath; respect to Figure 2. All the other temperatures were measured then the Omega datalogger was setup in order to measure with the thicker type. the temperature from each thermocouple every 10 s. Water was inserted into the thermos up to the previously defined E. Thermos maximum water level and the thermos would then be closed. A vacuum insulated thermos with nominal capacity of 0,5 L The temperatures would then be recorded until the temperature was used. To insert the thermocouples inside the thermos, with of the water inside would reach ambient temperature. the aim of monitoring and recording the water temperature and 2) External PCM layer: As a first approach, it was decided with the intention of creating the smallest heat loss possible, to drill 3 holes inside the first thermos layer and fill the cavity a small hole in the top of the cap was made.Additionally, all with PCM. This option would maintain the same amount of the internal thermoucouples were inserted in a silicon tube in water inside the thermos without increasing the volume. The total volume available in the cavity was of 0,180 L (179,6 g of order to increase adherence to the thermos hole and further ◦ reduce heat losses. water at 20 C). By melting the paraffin wax and inserting it in the cavity it was possible to fill the cavity with 127,5 g of PCM. F. Experimental setup and procedures For the first experiments, also thermocouples were inserted The various experimental setups used throughout this work in correspondence of the three holes (TA, TB and TC in are reported with diagrams outlining the thermocouples posi- Figure 2), but due to pressure build up inside the cavity these tions and type. In all the experiments in which hot water was thermocouples where removed and the container was sealed. used, a thermal bath of distilled water at 90 ◦C was employed. Additionally, since the isolating vacuum layer was now used The maximum useful quantity of PCM was calculated. to contain the PCM, experiments were run by wrapping the Taking into account only the sealed thermos and admitting an thermos with layers of granulated cork (k=0,045 W m−1 K−1 ideal scenario where there are no losses to the environment, [4]) to provide insulation and decrease the amount of heat LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 4 released by the thermos. Two different cork layers were latex containers allowed to put the PCM in practical compact, experimented. One layer was 1,2 mm thick, and the other 6 mm long and flexible containers with few grams of paraffin per thick. Figure 2 provides a schematic of this system, reporting container (around 10 g each, allowing for a more flexible system the position of the thermocouples used (all but TA, TB and and providing an easier way to modify the mass to PCM ratio. TC). 4) Cups: A simpler system than the thermos was also tested. Normal household cups were tested using the latex containers. The maximum amount of water that was possible to insert in the cup was of 300 g, and it was marked on the cup. With this setup it was possible to easily measure the effects of different water to PCM mass ratios. Figure 3 reports the position of the various thermocouples. The same latex containers described in the previous section were used.

Figure 3. Thermocouple position for the cup experiments. T1: 0,5 cm from the top surface of the water. T2: in the middle of the cup. T3: 0,5 cm from the bottom of the cup. T4: environment. Drawing not in scale.

Figure 2. Thermocouple position for the external PCM layer experiment. Each IV. EXPERIMENTAL RESULTS thermocouple is designated with the letter ‘T’ followed by an identification number or letter. T1: middle of outer thermos wall. T2: 1 cm from the top A. Paraffin characterization results surface of the water. T3: in the middle of the thermos. T4: 1 cm from the bottom of the thermos. T5: middle of teh inner wall. T6: environment. T7: The data obtained in this section allow the comparison of in the middle of the outer cork layer. TA, TB, TC: in contact with the PCM. the various thermo-physical properties between the two chosen Drawing not in scale. PCM. In addition, by knowing all of these data it is possible to create an accurate optimization model of the system in a 3) Internal PCM container: A different approach was also future work and compare the properties of the paraffin wax tested. For this experiments, an unmodified thermos was used, with the sodium acetate trihydrate. and the PCM was inserted in the main compartment. As a 1) DSC results: Figure 4 reports the DSC results, corres- downside, this created a significant decrease in the maximum ponding to the second heating cycle of the paraffin wax at amount of water that is possible to load in the thermos. a scanning rate of 10 ◦C min−1. The reported data are the To insert the paraffin wax in the thermos different methods ◦ onset melting temperature Tfus=55,84 C and the latent heat were used. Two measures of stainless steel cylinders were used: −1 of fusion ∆Hfus=182,4 J g . As it is possible to observe, two a low one (Cylinder 1) and a tall one (Cylinder 2). The small peaks are present. Although the higher one is the main one, one was 9,48 cm high with an external diameter of 2,8 cm and the second peak reports the presence of another paraffin with a wall thickness of 1 mm. The tall one had the same diameter a lower melting temperature in a relatively substantial amount. and wall thickness but it was 16,25 cm high. The amount of ◦ The peak solidifying temperature is found at Tfus=51,29 C PCM in each tube is reported in Table IV. Additionally, to with a similar enthalphy of solidification. test surface contact area influence, Cylinder 3, identical to 2) Density results: The calculated density results and their Cylinder 2 but with the addition of stainless steel wool inside associated errors are reported in Figure 5 on the following was experimented in order to increase the heat exchange area page. inside. The weight of the inserted wool was of 22,4 g With the 3) Viscosity: Figure 6 on the next page reports the calculated same purpose, Cylinder 4 16,5 cm high and with a diameter of kinetic viscosity results and their associated error. 0,8 m was created. To use the same internal volume of Cylinder 4) Thermal diffusivity: Table II reports the value of thermal 2, fourteen of these cylinders were used. With this solution diffusivity for each shot and the average. instead of aggregating all the PCM in one container, it was divided into multiple cylinders with approximately a four times increase in surface area. The amount of PCM inserted resulted B. Thermos results to be less than Cylinder 2 because of the numerous technical 1) The reference experiment results: Figure 7 reports the issues encountered while sealing each cylinder. temperature over time graph for a normal unmodified thermos As a second approach, a flexible latex container was used in of 0,5 L. From the graph, it is possible to note that the external order to easily modify the water mass to paraffin wax ratio. The wall maximum temperature (T1) is around 25 ◦C with an Dynamic viscosity Dynamic Viscosity Dynamic Viscosity Kinetic Viscosity Kinetic Kinetic Viscosity Dynamic Relative Error Temperature [°C ±1] [mPas] Relative Err [kg/ms] [m2/s] Viscosity Err Relative Err Viscosity [g/ms]

60 4,51 0,01 4,51E-03 5,04E-06 6E-08 0,01 4,51 0,01 0,05

70 3,33 0,01 3,33E-03 3,77E-06 5E-08 0,01 3,33 0,01 0,03

83 2,50 0,01 2,50E-03 2,87E-06 4E-08 0,01 2,50 0,01 0,03

Volume Mass Total Temperature [°C ±1] Volume [µL ±4] Mass [g] ± 10µg Total Err % Density [kg/m3]=[g/L] Density Err Relative Err Relative Err Relative Err

55 500 0,008 0,45197 2,21E-05 0,01 0,80 904 7,23 0,43578

60 500 0,008 0,44747 2,23E-05 0,01 0,80 895 7,16 0,44125

70 500 0,008 0,44125 2,27E-05 0,01 0,80 883 7,06 0,44747

83 500 0,008 0,43578 2,29E-05 0,01 0,80 872 6,97 0,45197

Kinetic Viscosity [m2/s]

920 6,00E-06

4,50E-06 905 ] ³ ⁻

3,00E-06 890

Density [kg m [kg Density 1,50E-06 875

0,00E+00 860 60 70 83 LORA0,8 – MSC228,1 SELECT455,4 MASTER682,7 THESIS, EXTENDED910 SUMMARY, JULY 2014 5

Kinetic Viscosity Error Temperature [°C] [mm 60 5,04 0,06 Table II 70 3,77 0,05 HERMAL DIFFUSIVITY RESULTS 83 2,87 0,04 T .

Temperature Thermal diffusivity [cm2 s−1] [◦C] Average Shot 1 Shot 2 Shot 3 Shot 4 Shot 5 6,00 98 0,7574 -- 0,7682 0,7207 0,7833

] 5,04 147 0,7294 0,7105 0,7637 0,7634 0,6978 0,7116 ¹ ⁻ 5,00 s

² 198 0,7085 0,6995 0,7057 0,7191 0,7046 0,7137 297 0,6564 0,6556 0,6622 0,6523 0,6383 0,6735

4,00 3,77 398 0,6342 0,6554 0,6741 0,6472 0,5818 0,6124

3,00 2,87 Kinetic Viscosity [mm Viscosity Kinetic

2,00 55 60 65 70 75 80 85 T [°C]

Density [kg/ Temperature [°C ±1] Density Err m3]

55 904 7,23

60 895 7,16 Figure70 4. DSC second883 heating7,06 cycle result. 83 872 6,97

920

910 904

] 900 ³ ⁻

895

890 883

Density [kg m 880 Figure 7. Temperature over time graph for the reference experiment using

872 the thermos. 870

860 50 55 60 65 70 75 80 85 and "T3 - Cork 6[mm" refer to the same modiﬁed thermos T [°C] with 1,2 mm and 6 mm of isolation respectively (for which Figure 5. Liquid parafﬁn wax: density results and associated error. granulated cork was used). Dynamic viscosity Dynamic Viscosity Dynamic Viscosity Kinetic Viscosity Kinetic Kinetic Viscosity Dynamic Relative Error Temperature [°C ±1] [mPas] Relative Err [kg/ms] [m2/s] Viscosity Err Relative Err Viscosity [g/ms] Table III summarizes the water to PCM mass ratio for each 60 4,51 0,01 4,51E-03 5,04E-06 6E-08 0,01 4,51 0,01 0,05 ◦ 16,00 70 3,33 0,01 3,33E-03 3,77E-06 5E-08 0,01 3,33 0,01experiment,0,03 the time needed to reach 60 C and the amount 83 2,50 0,01 2,50E-03 2,87E-06 4E-08◦ 0,01 2,50 0,01 0,03 ◦ internal12,00 water temperature that is around 85 C. This result of time that the water is kept at optimum conditions (60 C to ] ¹ ⁻

s ◦ ² highlights the very goodVolume isolationMass ofTotal the thermos vacuum layer. 50 C). 8,00 Temperature [°C ±1] Volume [µL ±4] Mass [g] ± 10µg Total Err % Density [kg/m3]=[g/L] Density Err Relative Err Relative Err Relative Err 5,04 55 500 0,0083,77 0,45197 2,21E-05 0,01 0,80 904 7,23 0,43578 4,00The results show that the thermos,2,87 with an external ambient 60 500 0,008 0,44747◦ 2,23E-05 0,01 0,80 895 7,16 0,44125 temperature70 of500 around0,008 200,44125C need2,27E-05 0,01 more0,80 than 88317 h to7,06 reach0,44747 Table III 0,00 83 500 0,008 0,43578 2,29E-05 0,01 0,80 872 6,97 0,45197 40 ◦C from the initial pouring temperature of approximately SUMMARY OF EXPERIMENTAL CONDITIONS AND RESULTS FOR THE Kinetic Viscosity [mm Viscosity Kinetic Kinetic Viscosity [m2/s] EXTERNAL PCM LAYER TRIALS. 90-4,00◦C. 920 6,00E-06 -8,00 ◦ opt opt If60 the timeframe70 in which the83 water is between 60 C to m m t ∆t Experiment w PCM Ratio ◦ T [°C] 4,50E-06 50 C905is taken in consideration as the range in which the liquid [g] [g] [min] [min] ] ³ ⁻

contained in the thermos is at optimum3,00E-06 drinking temperature T3 - No isolation 468,0 127,5 3,67 27,5 47,5 deﬁned890 as "Optimum Timeframe", than the smaller capacity T3 - Cork 1,2[mm] 468,0 127,5 3,67 35,0 47,5

Density [kg m [kg Density ◦ 1,50E-06 T3 - Cork 6[mm] 468,0 127,5 3,67 45,8 77,3 thermos875 reaches 60 C ("Time to Optimum") after 25 000 s and keeps the water in the optimum range for around 16 000 s. 0,00E+00 860 60 70 83 The best result is obtained by the most insulated thermos. It 2) External0,8 228,1 PCM455,4 layer results:682,7 910Figure 8 shows the comparison of the results for three different experiments: "T3 - No is possible to observe that the decay of the temperature around Kinetic Viscosity Error Temperature [°C] [mm the melting point of the PCM is smoothed out by the release of isolation"60 refers to5,04 the the0,06 modiﬁed thermos with the PCM inserted70 in the vacuum3,77 0,05 gap, whereas "T3 - Cork 1,2[mm" its latent heat energy. As the optimum timeframe is increased, 83 2,87 0,04 also the Time to Optimum grows with the addition of insulation. This is an unwanted behavior, since it is preferable for the 6,00 drink to be readily available. The thinner layer of cork has

] 5,04 ¹ ⁻ 5,00 almost no effect on the optimum timeframe, but it increases s ² of around 27 % the Time to Optimum. Although these initial 4,00 3,77 results are promising, they are in no way comparable to the

3,00 2,87 reference experiments reported in the previous section. All of Kinetic Viscosity [mm Viscosity Kinetic the obtained results (Time to Optimum and optimum timeframe) 2,00 55 60 65 70 75 80 85 are an order of magnitude lower than the previous result. The T [°C] main reason for this is the extreme losses that the thermos has. Density [kg/ Temperature [°C ±1] Density Err m3] Figure 8 compares the results for the three different situations. Figure 6. Liquid55 parafﬁn904 wax:7,23 kinetic viscosity calculated results and associated error. 60 895 7,16 By performing an analysis and comparison of the external wall 70 883 7,06

83 872 6,97

920

910 904

900 ]

³ 895 ⁻

890

883 Density [kg m 880

872 870

860 50 55 60 65 70 75 80 85 T [°C]

16,00

12,00 ] ¹ ⁻ s ² 8,00 5,04 3,77 4,00 2,87

0,00

Kinetic Viscosity [mm Viscosity Kinetic -4,00

-8,00 60 70 83 T [°C] LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 6 temperature (T1 and T7) with the same temperature of the reference experiment of the 0,5 L thermos it is possible to observe a signiﬁcant increase in maximum temperature. In this case, the external wall reaches almost 60 ◦C. Although this effect is highly mitigated in the experiment with the better insulated thermos, which reports an external insulation temperature (T7) lower than 40 ◦C, the losses are still too high to create a comparable result with the reference experiment. In addition, due to the extreme temperature losses, not all PCM melts. TC and TA in fact, report peak temperatures lower Figure 9. Temperature over time comparison graph for all the internal PCM than 50 ◦C. container experiments; referred to the water thermocouple in the middle of the thermos (T3). Table IV reports each Cylinder type difference. For "Thermos", the reference thermos experiment data was used.

wool did not bring any benefit, it only marginally lowered the amount of PCM introduced in the cylinder. Especially for Cylinder 4, the moment in which the cooling temperature changes trend for the release of heat from the PCM is more pronounced. This is primarily due to the increased surface area used in this experiment, which allows all the PCM contained in each tube to receive heat from the hot water more evenly. In addition, although Cylinder 4 has a lower mass ratio, it Figure 8. Temperature over time comparison graph for the external PCM layer experiment. performs marginally better than all the other cylinders types, possibly for the same reason. The losses with this type of solution create a system that All the different solutions employing a PCM perform better does not improve a normal unmodified thermos. The vacuum than the reference experiment; it is therefore necessary to use layer creates a very effective isolation that is not possible a mass ratio that is closer to the minimum limit calculated in to replicate unless using expensive isolating materials and/or the previous chapter in order to maximize the performance. increasing excessively the insulating radius. The "Latex" and "SAT" experiments reported in Figure 9 focus It is possible to observe some initial signs of the PCM latent exactly on this point. "Latex" refers to the use of the latex heat storage effect, which are promising, but a solution in containers. Since these containers were modular and of smaller which the PCM is in contact only with the water is probably sizes compared with Cylinder 1, it was possible to insert more preferable. Another possible solution that can be explored is PCM inside the thermos. "SAT" refers to the solution of sodium the use of a double wall thermos, with the first internal wall acetate trihydrate, packed in a double wall plastic bag with filled with PCM and the external one that provides an isolating the correct proportion of nucleating agent as reported in the vacuum gap. previous chapter. The water to PCM ratio used is lower than 3) Internal PCM container results: Table IV reports a the minimum ratio calculated in the previous chapter, in order summary of the various results obtained using the thermos to use a ratio similar to the "Latex" experiment. Both of these for each cylinder type. Figure 9 reports a graph that compares solutions outperform any other attempt previously made. The all the various solutions. amount of water introduced is higher for the "SAT" experiment, since sodium acetate trihydrate has an higher density. For both Table IV SUMMARY OF EXPERIMENTAL RESULTS FOR THE INTERNAL PCM experiments the Time to Optimum performance is notably low, CONTAINERS."THERMOS" REFERS TO THE REFERENCE EXPERIMENT DATA. around 25 min instead of almost 7 h that the thermos without PCM needs. In addition, both increase the Optimum Timeframe m m topt ∆topt Experiment w PCM Ratio to approximately 6,5 h instead of the 4,4 h of the reference [g] [g] [h] [h] thermos. As a drawback, the amount of water that is possible Thermos 0,5 L 468,0 0 - 6,94 4,44 to insert is significantly decreased of 64 % for the "Latex" Cylinder 1 411,2 37,14 11,07 6,18 6,04 Cylinder 2 376 65,43 5,75 4,31 6,11 container and of 49 % for the "SAT" container. This result Cylinder 3 376 64,33 5,85 4,38 5,63 translates in the need of a bigger volume to transport the same 14 · Cylinder 4 346,7 56 6,2 4,03 6,38 amount of water. Latex 285,2 113,8 2,51 0,35 6,74 SAT 314,7 124,1 2,54 0,49 6,46 Considering the parameters reported in Table IV, the "Latex" container performs slightly better than the "SAT" one, since An increase in the water to PCM mass ratio also increases it uses a PCM disposition similar to Cylinder 4, i.e. it has an the amount of time during which water stays in the Optimum overall higher surface area. Analyzing the curve of the "SAT" Timeframe. Although Cylinder 3 is very similar to Cylinder it is possible to observe how the use of an higher melting point 2, it performs slightly worse. The insertion of stainless steel PCM keeps the water warmer in confront with the "Latex" LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 7 container that uses the paraffin wax with a lower solidifying V. PLANNINGTHEBUSINESS point. A business analysis is needed to determine the market for a 4) Cup results: Table V reports the mass ratio, the Time to potential product and to prove that it is also financially feasible. Optimum drinking temperature and the Optimum Timeframe for different mass ratios, with a reference cup filled only with A. The need and customer water. The need for food and beverages is one of the constant variables in our lives. Since the introduction of the vacuum Table V container at the the beginning of the 20th century [12] there has SUMMARY OF EXPERIMENTAL RESULTS FOR THE CUP EXPERIMENTS. been little to no evolution from the initial design. As a result m m topt ∆topt of this lack of innovations, users are now widely accustomed Experiment w PCM Ratio [g] [g] [min] [min] not to be able to control the temperature of their hot beverage T1 - Water 300,0 0 - 14,5 13,2 and take it as it is, so that now it is widely accepted for a T2 - Low Ratio 200,1 68,0 2,94 2,7 16,5 drink to be either too hot or too cold. Two start ups which T2 - Middle Ratio 210,5 56,5 3,73 1,0 12,0 T2 - High Ratio 229,7 46,4 4,95 8,2 14,8 have already been trying to solve this problem by using a PCM were met with high enthusiasm on the popular crowdfunding website Kickstarter [7], [13]. Although all the experiments overall performed better then There has also been an increased awareness of the general the reference one containing only water, in this case there is public towards sustainability and sustainable solutions. In this no clear mass ratio that performs better than the others. The direction it is important to underline that the system presented Middle Ratio contains an amount of PCM that is probably the with this work does not need any additional external energy closest amount to the optimum ratio, which allows it to cool input. The introduction of PCM containers in this market the temperature of the water to an optimum value in around segment would simply use heat that is currently wasted to the 1 min, faster than any other option. The downside is that it environment in a better and more efficient way. keeps the water at optimum conditions for the least amount of time. As a result of this situation, the Low Ratio seems to B. Business Model have the overall best performance, since it can cool the liquid By developing a functional design it is possible to build at optimum temperature in less than 3 min and it can keep it a prototype that shows how the system works, providing a for the longest amount of time. basis to showcase the potential of the project to prospective If instead what is looked for is the overall longer period above investors. The search for partner manufacturing facilities ◦ 50 C, than the unmodified cup has a clear advantage. It can be needs to go along with the search for capital seed money, possible to introduce a small quantity of PCM, such as in the in order to optimize production process together with the High Ratio experiment, to almost halve the time needed to reach design finalization. In case of lack of capital seed investments, optimum temperature and increase the Optimum Timeframe crowdfounding the project can be considered as a viable option. for around 2 min, but the downside is a 23,4 % lower amount To pursue the described business strategy, a possible Business of water in the cup. Model, shown in figurename 11 that an Técnico Lisboa (IST) spin-off company could use is reported.

Figure 10. Temperature over time complete graph for the Low Ratio experiment

There are very high water temperature stratifications between ◦ the top (T1) and the bottom (T3) water thermocouple (>20 C). Figure 11. Proposed Business Model. On the contrary, in the experiment with the cup only filled with water, this behavior is not encountered; in fact, there is no significant stratification. This behavior is probably due to C. Analysis of the macro environment the PCM, which creates high convection currents as soon as it A political, economic, social and technological (PEST) gets in contact with hot water and starts absorbing heat. As analysis is necessary to properly evaluate the influence of heat is lost to the environment, this effect becomes weaker and the all the conditions and of the environment that surrounds the phenomena decreases in magnitude.1 the company. LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 8

1) Political factors: The Kickstarter page also proposes several temperature over a) Food Safety: The main document of the European time graphs, which show that with the Temperfect mug coffee Union (EU) regarding materials that can come in contact is cooled to 60 ◦C from 78 ◦C in around 225 s, and keeps the with food intended for human consumption is contained in liquid at Optimum Temperature for approximately 9500 s. The ‘Guidelines on metals and alloys used as food contact materials’ project financing finished in January 2014, reaching a total [14]. Although each country has slightly different legislation of 269 271 $ and 4903 people that backed the project. The (with France and Italy having the more stringent laws), any amount of water the mug can hold is almost 0,5 L, exactly as kind of stainless steel is generally accepted. The EU also the thermos used in this work, but the Temperfect Mug does strictly regulates physical hazards, which must be prevented, not provide a leak tight seal. eliminated or reduced to an acceptable level [15]. In order to achieve this, all parties involved should use the hazard analysis E. Case Study: Thermos design critical control point (HACCP) methodology, in addition to By performing an analysis of all the results achieved during quality control and management systems such as ISO 9000 this work and by what competitors have been producing and [15]. selling, a reference design was created, aiming to tackle all the 2) Economic factors: Since 2008, the EU along with the negative aspects of each different solution while maximizing greater part of the rest of the world entered an heavy recession, performance. The designed thermos and PCM vessel applies as data from Eurostat [16] about gross domestic product (GDP) principles that can also be exported to a smaller travel or desk growth rates confirms. Despite this fact, drinking liquids is coffee mug (200 mL to 330 mL). a necessity, and this events did not influence beverages. In a) Esternal container: The design comprises a liquid fact, worldwide tea and coffee production increased during this container made of stainless steel isolated by a vacuum gap. period [17]. As a consequence, it is possible to affirm that the The bottom part of the liquid container has the size to fit a market is not in a shrinking phase. normal car cup-holder. The container diameter then increases as 3) Social factors: People are creatures of habit. Once they it reaches the top. This feature allows for an easier access to the have developed a pattern in behavior that proved to be effective internal compartment of the thermos, especially for cleaning. they stick to it, preferring what they know to the uncertainty b) Lid: The vacuum thermos is closed by a spill proof of the unknown [18]. In particular, a product may be better lid. This feature provides the flexibility necessary to safely accepted if its use does not require a significantly bigger effort transport the thermos without worrying about spills. The lid than the one they are used to do. is kept in place by a locking mechanism, e.g. a screw thread, 4) Technical factors: Innovations that may affect the opera- and it is possible to completely remove the lid to allow access tions of the industry and the market are analyzed. Nowadays to the internal compartment of the liquid container. as it is possible to observe, by examining the product lineup c) PCM: Confronting the two PCM chosen for this work, of the main producers, that differentiation is created mainly sodium acetate trihydrate is expected to perform better than with product design. the paraffin wax tested. Despite this, futher studies should be conduced and other PCM should also be analyzed and kept into consideration for future reference. D. Analysis of the micro environment d) PCM container: The container is the most innovative 1) Thermos L.L.C.: Undiscussed leader of the market and and important part of the system. It is made of an hollow holder of the first patent regarding the vacuum container cylindrical tube filled with slightly pressurized PCM. The technology; it is also the company that invented the Pyrex pressure is needed to ensure that all the container is filled glass vacuum container and the stainless steel one. Although with the PCM and all the volume is used. The cylinders need currently they do not offer PCM enhanced vacuum containers, to be correctly sized by an optimization study, with water to they have the size, funds and experience to quickly break into PCM mass ratio and contact surface area as main parameters. it. Naturally, results are dependent from the chosen PCM. 2) Coffee Joulies: They are a relatively new and small These cylinders can be rendered magnetic with the addition player in the market, but they constitute an unique case of magnets in the hollow cavity. By doing this they can be because at the moment they are the only ones offering a stacked and kept together inside the thermos. By magnetically PCM enhanced container. Their homonymous product was attaching them also to the lid or internal bottom part of the first launched on Kickstarter in 2011. The system is comprised liquid container they would not move when transported, while of a stainless steel vacuum flask and 5 coffee bean shaped at the same time granting a better optimized contact with water stainless steel PCM containers. The product works at his best since the cylinder(s) are surrounded by it. The use of more than when the PCM container beans are preheated and placed in one cylinder offers flexibility, i.e. the possibility for the system the vacuum container along with tea or coffee. Although the to work with containers of smaller dimensions, such as travel project received a huge success, collecting 306 944 $ from 4818 mugs. The magnets also allow for easier transportation and founders, costumers reviews once the product was shipped are cleaning. A detachable system gives the possibility to choose highly discordant. between a system that keeps drinks cold or hot by simply 3) Joeveo: The product, called Temperfect Mug, is still in swapping the PCM cylinders. In all cases, the lid or the bottom the final development stages. The new product is made of a of the liquid container need to be magnetic for the cylinder to vacuum insulated mug with an internal layer filled with PCM. be attached to the container. REFERENCES 9

1) Preliminary financial analysis: By analyzing the price feasible. To obtain a more accurate production costs, a future points of Joeveo and Coffee Joulies it is proven that people financial analysis will need to be performed once it is possible are willing to pay between the 40 $ of the Joeveo system and to have more precise data about the costs of producing the PCM the 100 $ of the Coffee Joulies ones. A bill of materials can container and about the real cost structure of the company. be made for the construction of one prototype (1 vacuum flask and 5 PCM containers) as shown in Table VI. The prices VI.CONCLUSIONS reported are intended for private users, i.e. they are common The following summarized conclusions can be drawn: retail prices. The number and price of each PCM container • The patent research about PCM enhanced thermos, mugs is difficult to calculate because of a lack of manufacturing and dishes shows that there is a concrete interest about costs data. A worst case scenario in which one empty PCM this sector. container has a cost similar to a single Coffee Joulies bean is • DSC was used to determine the paraffin wax melting considered. (55,84 ◦C) and solidifying point (51,29 ◦C), as well as the latent heat of fusion (182,4 J g−1). Table VI ESTIMATED PROTOTYPE COST. • For water and ambient temperature conditions similar to the experimental ones, the minimum water to PCM Material Price Unit Amount Reference mass ratio was identified in an ideal scenario, taking into Thermos 10–25 $/piece 1 [19] account the specific and latent heat of each PCM. The PCM 27 $/kg 183 g [20] mass ratio calculated with this method amounts to 2,0 Magnets 0,01–0,2 $/piece 11 [21] PCM container 7 $/piece 5 [7] for the paraffin wax and to 2,7 for the sodium acetate trihydrate. Total 50,1–67,1 $/system • Results showed that higher surface areas (4 to 5 times bigger) perform slightly better than lower ones (4 to 5 % To estimate a production run of, for example, 5000 units, increase of ∆topt respectively). similar to the number of customers for Joeveo during their • In comparison with a standard vacuum thermos, using a crowdfunding campaign, economy of scale prices can be PCM can decrease topt up to 95 % faster and keep ∆topt applied. Table VII reports the cost of the various materials if up to 50 % longer. bulk quantities are bought. As in the previous case it is difficult • The denser and with higher latent heat of fusion sodium to estimate the cost of producing the PCM container. For this acetate trihydrate performs better than the paraffin wax. reason, the previously used price is taken in consideration and The material, which has a melting temperature closer reduced of taxes and gross profit margin. Portuguese value to 60 ◦C, overall keeps the liquid warmer during the added tax (VAT=23 %) was considered and a gross profit margin Optimum Timeframe. of 32 % is assumed based on the data rof Butler Consultants • The use of PCM materials in normal kitchen cups is [22] for an equipment manufacturer. discordant. Unless for specific situations, normal cups perform better. The lack of insulation and associated Table VII thermal losses are too high to allow the PCM to bring ESTIMATED COST FOR A 5000 UNITSPRODUCTIONRUN. significant benefits. Material Price Unit Amount Reference • The main competitors were identified, namely in Thermos Thermos 2–3,5 $/piece 5000 [23] L.L.C. [19], Cofee Joulies [7] and Joeveo [13]. The PCM 600–650 $/Mg 915 kg [24] latter two have already proposed PCM enhanced thermos Magnets 0,001–0,5 $/piece 55 000 [25] solutions. PCM container 3,66 $/piece 25 000 [7] • A bill of materials analysis for the proposed design, which Total 102 104–137 095 $/5000 system is intended to improve the current ones, was made. A Unit Total 20,4–27,4 $/system comparison with current marketed systems costs proves the proposed design financially feasible. With bigger production runs it is possible to produce each thermos for a lower price. As in the previous case, ACKNOWLEDGMENT the biggest uncertainty is the cost of the PCM container. Similar considerations can also be applied: manufacturing and The author would like to thank all the Professors and development costs, taxes and profit margins are still missing. researchers of Técnico Lisboa that aided the development of If the same VAT and gross profit margin values are applied to this work, Dott. Rui Costa Neto in particular. the worst case scenarios, than a possible final retail cost for the product would be of 109 $ for the prototype and of 45 $ for REFERENCES large scale production run. Comparing these results with the [1] A. Cruz and B. G. Green, ‘Thermal stimulation of taste’, Coffee Joulies and Joeveo systems prices, the prototype cost is Nature, vol. 403, 24th Feb. 2000. only 9 $ higher than the final Coffee Joulies price, and the scale [2] K. Talavera, K. Yasumatsu, T. Voets, G. Droogmans, N. production run price per unit is just 5 $ more than the Joeveo Shigemura, Y. Ninomiya, R. F. Margolskee and B. Nilius, system. Taking into account that these are estimated worst ‘Heat activation of trpm5 underlies thermal sensitivity case scenario prices, the product can be considered financially of sweet taste’, Nature, vol. 438, 15th Dec. 2005. LORA – MSC SELECT MASTER THESIS, EXTENDED SUMMARY, JULY 2014 10

[3] P. Mony, T. Tokar, P. Pang, A. Fiegel, J.-F. Meullenet and [20] P212121. (2014). Sodium acetate trihydrate 1 kg, [On- H.-S. Seo, ‘Temperature of served water can modulate line]. Available: http://store.p212121.com/Sodium- sensory perception and acceptance of food’, Food acetate-trihydrate/ (visited on 01/07/2014). Quality and Preference, vol. 28, no. 2, pp. 449–455, [21] Emovendo. (2014). 30 neodymium magnets, [Online]. 2013. Available: www.amazon.com/Neodymium- Magnets- [4] F. P. Incropera, Fundamentals of Heat and Mass Transfer. inch - Disc - N48 / dp / B001KUWM3C/ (visited on John Wiley & Sons, 2006. 05/07/2014). [5] H. Mehling and L. Cabeza, Heat and cold storage [22] Butler Consultants. (2014). Free industry statistics - with PCM: An up to date introduction into basics and sorted by highest gross margin, [Online]. Available: applications, ser. Heat and Mass Transfer. Springer, http://research.financial-projections.com/IndustryStats- 2008. GrossMargin (visited on 05/07/2014). [6] A. Sharma, V. Tyagi, C. Chen and D. Buddhi, ‘Review [23] Zhejiang Rongxuan Industrial Co., Ltd. (2014). Acciaio on thermal energy storage with phase change materials inox novità riciclato moderno caffè tazza thermos fiasco, and applications’, Renewable and Sustainable Energy [Online]. Available: http://italian.alibaba.com/product- Reviews, vol. 13, no. 2, pp. 318–345, 2009. gs / stainless - steel - novelty - recycled - modern - coffee - [7] Cofee Joulies. (). Coffee joulies, Your coffe, just right, mug-thermos-flask-1805146222.html?s=p (visited on [Online]. Available: www . joulies . com (visited on 05/07/2014). 26/06/2013). [24] Foodchem International Corporation. (2014). Acetato [8] F. Brown and K. R. Diller, ‘Calculating the optimum di sodio triidrato, [Online]. Available: http : / / italian . temperature for serving hot beverages’, Burns, vol. 34, alibaba . com / product - gs / sodium - acetate - trihydrate - no. 5, pp. 648–654, 2008. 1631448628.html?s=p (visited on 05/07/2014). [9] G. A. Lane, Latent heat materials. Boca Raton, Florida: [25] Sky Magnetech (Ningbo) Co., Ltd. (2014). Neodimio CRC Press, 1983, vol. 1. calamita, [Online]. Available: http://italian.alibaba.com/ [10] M. Keinänen, ‘Latent heat recovery from supercooled product-gs/neodynium-magnet-1769327501.html?s=p sodium acetate trihydrate using a brush heat exchanger’, (visited on 05/07/2014). Master’s thesis, Helsinki university of technology, 2007. [11] J. Mao, J. Li, G. Peng and J. Li, ‘A selection and NOMENCLATURE optimization experimental study of additives to thermal ρ Density energy storage material sodium acetate trihydrate’, in ∆Hfus Latent heat of fusion Energy and Environment Technology, 2009. ICEET ’09. mPCM Mass of PCM International Conference on, vol. 1, Oct. 2009, pp. 14– mw Mass of water 17. ∆topt Optimum Time Frame: time that a liquid stays in [12] American Thermos Bottle Co, ‘Double-walled vessel optimum conditions (60 to 50 ◦C) with a space for a vacuum between the walls’, Patent cp Specific heat capacity at constant pressure US872795 A, 3rd Dec. 1907. cp-PCM PCM specific heat capacity at constant pressure [13] Joeveo. (2014). Temperfect mug, [Online]. Available: cp-w Water specific heat capacity at constant pressure http://joeveo.com (visited on 11/06/2014). Tfus Temperature of solid-liquid transition [14] Council of Europe, ‘Guidelines on metals and alloys k Thermal conductivity used as food contact materials, Technical document’, α Thermal diffusivity Strasbourg, Tech. Rep., 13th Feb. 2002. topt Time to Optimum: time that a liquid needs to reach [15] European Parliament and Council of the European Union, 60 ◦C Regulation (ec) no 852/2004 on the hygiene of foodstuffs, Netherlands, 2004. ACRONYMS [16] Eurostat. (2014). Real gdp growth rate - volume, Percent- DSC Differential Scanning Calorimeter age change on previous year, [Online]. Available: http: DSP Disodium Hydrogen Phosphate Dodecahydrate //epp.eurostat.ec.europa.eu/tgm/table.do?tab=table& EU European Union init = 1 & plugin = 1 & language = en & pcode = tec00115 GDP Gross Domestic Product (visited on 01/07/2014). HACCP Hazard Analysis Critical Control Point [17] Food and Agriculture Organization of the United Nations. IST Técnico Lisboa (2014). Production growth rates by region, 2001 - LFA Laser Flash Analysis 2012, [Online]. Available: faostat.fao.org (visited on PCM Phase Change Material 10/06/2014). PEST Political, Economic, Social And Technological [18] M. C. Gonzalez, C. A. Hidalgo and A.-L. Barabasi, ‘Understanding individual human mobility patterns’, Nature, vol. 453, pp. 779–782, 2008. [19] Thermos L.L.C. (2014). Products, Beverage bottles, [Online]. Available: www.thermos.com/product_catalog. aspx?CatCode=BEVG (visited on 11/06/2014).